Terrestrial plants are exposed to higher CO₂ concentrations today than at any time over the last 650,000 years. Plant responses to CO₂ can be categorized into three groups: (1) short-term response (seconds to minutes), (2) acclimation (days to a few years), and (3) adaptation (over generations). While short-term responses and acclimation are relatively well understood, adaptation to high CO₂ is poorly understood. Study of adaption to high CO₂ is difficult because experiments cannot be designed on the appropriate timescale required to observe evolution. Natural CO₂ springs are unique and attractive ecosystems to study evolutionary responses to high CO₂ because vegetation around the spring has been exposed to high CO₂ over generations. We have found several natural CO₂ springs that were useful for this purpose. A common garden experiment, taking plants from these CO₂ springs and nearby control sites, showed that some spring-plants had genetically lower stomatal conductance, higher water use efficiency, lower starch accumulation and higher shoot-root ratio. Micro-satellite analysis revealed relatively large genetic differentiation across the CO₂ gradient. These results support the idea that atmospheric CO₂ elevation can act as a selective agent on some plant traits in natural plant communities.